of Hydrocarbons and their Mixtures. 123 



is crucial with respect to this method. The percentage of 

 ethylene in the last but one is nearly twice that in the last, 

 yet the C.P. differs but little. The explanation is that here 

 the carbon of the ethylene is added to the considerable quan- 

 tity of carbon already present in the form of marsh-gas. 



I now come to a point of some difficulty. If this straight 

 line represents the behaviour of mixtures of ethylene and 

 marsh-gas as well as it appears to do, it should also represent 

 the case of marsh-gas itself. But seeking * the point in the 

 line corresponding to a? = '25, we find ?/ = 12'4. This indi- 

 cates what we may call a theoretical candle-power of 12'4 for 

 marsh-gas. 



Now it is certain that marsh-gas does possess illuminating- 

 power. P. Frankland states that marsh-gas determined in 

 the Referee's burner gives a power of " several candles " f. 

 Also Wright has determined its lighting-power, which he 

 makes about 5 candles (Chem. Journ. 1885, p. 202). 



But admitting that the gas has some illuminating-power, 

 it will not he generally thought that this can amount to 12 

 candles. 



There are two views which may be taken. The one, that 

 the straight line fails here to represent the properties of the 

 mixture when the ethylene vanishes. This is quite possible. 

 If so, a curved tail will have to be added to represent the 

 pure marsh-gas, when it has been determined exactly what 

 the candle-power is that it can be supposed to have. 



The other view, to which I incline, is that our methods are 

 not adapted for the measurement of illuminating-powers of 

 small intensity ; and that for the present we must treat such 

 cases with reserve. 



Turn back for a moment to the series of mixtures of ethy- 

 lene and hydrogen. The lowest point falls considerably 

 below a line drawn through the two next above. Now it 

 may be that the illuminating value shown by this lowest 

 point is too low for the reason above suggested, the inade- 

 quacy of the methods for very low powers. If so, the real 

 course of the hydrogen mixtures may be substantially different 

 from that shown. 



* The equation was formed to the line passing through the first and 

 last but one of the mixture points. This line gave 



C.P. 



Ethylene 65"5 



Marsh-gas 12'4 



I do not enter at length into the discussion of the ethylene values ; but 

 this was regarded as the most important of the values employed, 

 t Chem. Journ. 1884, p. 39. 



